What is a Good 3-Bet Percentage in Poker?

What the stat counts

3-bet percentage is the share of hands where you re-raised when facing exactly one raise preflop. The denominator is the part people forget: it is not every hand you play, it is only the hands where exactly one player had raised when the action reached you. That makes it a conditional frequency, and it has two consequences. First, a 3-bet percentage cannot be compared with a VPIP or a PFR, which are shares of all hands dealt. Second, it arrives slowly, because you have to be dealt a hand, somebody has to open, and nobody else can have re-raised first. Squeezes (a 3-bet with a caller already in) are usually counted separately by trackers, and they are a different decision, so it is worth knowing which your HUD is showing you.

Where the band comes from, and how often to 3-bet

The 5-10% guideline is a population figure, and our range library lets you check what a sound strategy actually produces. Counted in combinations of the 1,326 starting hands, the curated 3-bet ranges are:

• •Against an under-the-gun open: 2.6%. A range that tight opens tight, so there is nothing to attack and your 3-bet is pure value.
• •Against a cutoff open: 4.8%. Value plus a slice of hands that can keep barrelling.
• •Against a button open: 8.3%. The button opens 40.0% of hands against the 9.8% under the gun opens with, and the wider the opening range, the more of your own range profitably re-raises it.
• •Squeezing over an open and a caller from the button: 5.1%. Dead money in the pot, two capped ranges to fold out, but two players who can call.

The average is a consequence, not a target

Weight each of those ranges by how often that seat is the one who opened (button opens are the commonest at the table, early opens the rarest), and the overall frequency comes out at 5.8%, which sits in the lower half of the 5-10% band. Two things follow. A player at 5% is not automatically a nit; they may simply face a table of tight openers. And a player at 10% is not automatically aggressive; if the figure is built from 3-betting early opens rather than late ones, it is the same number describing a worse strategy. The model behind that 5.8% assumes every seat opens the library's range and nobody limps, which is exactly where it stops being true: a limp-heavy table gives you iso-raises instead of 3-bets, and a table of maniacs gives you far more of both.

What sits outside the band

Under the band the cost is not the pot you failed to win: it is the pots you play instead, flat-calling out of position with a range you have advertised as capped. Over the band the cost depends entirely on which hands you added. Weak suited aces and suited connectors against a late open have a plan for the flop; offsuit broadways against an early open have none, and they fold to the 4-bet that the strongest opening ranges are built to make. The practical test is not the number at all: it is whether you can name the flops you keep betting after your 3-bet gets called, and whether your 3-bets fold out anything at all.

How long before your own figure means anything

You need about 430 chances to 3-bet before a 95% interval on your own frequency is narrower than the 2.5-point half-width of the band. The opening model says somebody has raised before you act on 34.9% of hands, so that is roughly 1,200 hands, several times what VPIP and PFR need, and that is for the single average. A per-opener breakdown, which is the only version worth acting on, needs a multiple of it again. Until then, treat an opponent's 3-bet percentage over a few hundred hands as a hint about their style and nothing more.

Check your own figure

The free leak report computes your 3-bet frequency from a hand history and marks it against this band, alongside the fold-to-3-bet figure that tells you whether your own opens are being attacked. The range library holds the three curated 3-bet ranges quoted above if you want to compare hand by hand rather than frequency by frequency.